A physical model was utilized to show that the neural system can memorize a target position and is able to cause motor and sensory events that move the arm to a target with more accuracy. However, this cannot indicate in which coordinates the necessary computations are carried out. Turning off the lights causes the error to increase which is accomplished by cutting off one feedback path. The geometrical properties of arm kinematics and the properties of the kinesthetic and visual sensorial (...) systems should be better known before inferences about higher levels of processing can be drawn. (shrink)

Segregated neural circuits that effect particular domain-specific behaviors can be differentiated from neuroanatomical structures implicated in many different aspects of behavior. The basal ganglionic components of circuits regulating nonlinguistic motor behavior, speech, and syntax all function in a similar manner. Hence, it is unlikely that special properties and evolutionary mechanisms are associated with the neural bases of human language.

Neurobiological models of language need a level of analysis that can account for the typical range of language phenomena. Because linguistically motivated models have been successful in explaining numerous language properties, it is premature to dismiss them as biologically irrelevant. Models attempting to unify neurobiology and linguistics need to be sensitive to both sources of evidence.

Müller misconstrues autonomy to mean strict locality of brain function, something quite different from the functional autonomy that linguists claim. Similarly, he misperceives the interaction of learned and innate components hypothesized in current generative models. Evidence from sign languages, Creole languages, and neurological studies of rare forms of aphasia also argues against his conclusions.

The emphasis in the connectionist sentence‐processing literature on distributed representation and emergence of grammar from such systems can easily obscure the often close relations between connectionist and symbolist systems. This paper argues that the Simple Recurrent Network (SRN) models proposed by Jordan (1989) and Elman (1990) are more directly related to stochastic Part‐of‐Speech (POS) Taggers than to parsers or grammars as such, while auto‐associative memory models of the kind pioneered by Longuet–Higgins, Willshaw, Pollack and others may be useful for grammar (...) induction from a network‐based conceptual structure as well as for structure‐building. These observations suggest some interesting new directions for specifically connectionist sentence processing research, including more efficient representations for finite state machines, and acquisition devices based on a distinctively connectionist basis for grounded symbolist conceptual structure. (shrink)

Conceptual features differ in how mentally tranformable they are. A robin that does not eat is harder to imagine than a robin that does not chirp. We argue that features are immutable to the extent that they are central in a network of dependency relations. The immutability of a feature reflects how much the internal structure of a concept depends on that feature; i.e., how much the feature contributes to the concept's coherence. Complementarily, mutability reflects the aspects in which a (...) concept is flexible. We show that features can be reliably ordered according to their mutability using tasks that require people to conceive of objects missing a feature, and that mutability (conceptual centrality) can be distinguished from category centrality and from diagnosticity and salience. We test a model of mutability based on asymmetric, unlabeled, pairwise dependency relations. With no free parameters, the model provides reasonable fits to data. Qualitative tests of the model show that mutability judgments are unaffected by the type of dependency relation and that dependency structure influences judgments of variability. (shrink)

A possible relation between Derrida's deconstruction of metaphysics and connectionism is explored by considering diffeacuterance in neural nets terms. First diffeacuterance, as the crossing of Saussurian difference and Freudian deferral, is modeled and then the fuller 'sheaf of diffeacuterance is taken up. The metaphysically conceived brain has two versions: in the traditional computational version the brain processes information like a computer and in the connectionist version the brain computes input vector to output vector transformations non-symbolically. The 'deconstructed brain' neither processes (...) information nor computes functions but is spontaneously economical. (shrink)

One of the most influential arguments against the claim that computers can think is that while our intentionality is intrinsic, that of computers is derived: it is parasitic on the intentionality of the programmer who designed the computer-program. Daniel Dennett chose a surprising strategy for arguing against this asymmetry: instead of denying that the intentionality of computers is derived, he endeavours to argue that human intentionality is derived too. I intend to examine that biological plausibility of Dennett’s suggestion and show (...) that Dennett’s argument for the claim that human intentionality is derived because it was designed by natural selection is based on the misunderstanding of how natural selection works. (shrink)

This paper introduces a special issue of Cognitive Science initiated on the 25th anniversary of the publication of Parallel Distributed Processing (PDP), a two-volume work that introduced the use of neural network models as vehicles for understanding cognition. The collection surveys the core commitments of the PDP framework, the key issues the framework has addressed, and the debates the framework has spawned, and presents viewpoints on the current status of these issues. The articles focus on both historical roots and contemporary (...) developments in learning, optimality theory, perception, memory, language, conceptual knowledge, cognitive control, and consciousness. Here we consider the approach more generally, reviewing the original motivations, the resulting framework, and the central tenets of the underlying theory. We then evaluate the impact of PDP both on the field at large and within specific subdomains of cognitive science and consider the current role of PDP models within the broader landscape of contemporary theoretical frameworks in cognitive science. Looking to the future, we consider the implications for cognitive science of the recent success of machine learning systems called “deep networks”—systems that build on key ideas presented in the PDP volumes. (shrink)

We present a model for several early stages of the sensorimotor transformations involved in targeted arm movement. In psychophysical experiments, human subjects pointed to the remembered locations of randomly placed targets in three-dimensional space. They made consistent errors in distance, and from these errors stages in the sensorimotor transformation were deduced. When subjects attempted to move the right index finger to a virtual target they consistently undershot the distance of the more distal targets. Other experiments indicated that the error was (...) in the sensorimotor transformation rather than in the perception of distance. The error was most consistent when evaluated using a spherical coordinate system based at the right shoulder, indicating that the neural representation of target parameters is transformed from a retinocentric representation to a shoulder-centered representation. According to the model, the error in distance results from the neural implementation of a linear approximation in the algorithm to transform shoulder-centered target parameters into a set of arm orientations appropriate for placing the finger on the target. The transformation to final arm orientations places visually derived information into a frame of reference where it can readily be combined with kinesthetically derived information about initial arm orientations. The combination of these representations of initial and final arm orientations could give rise to the representation of movement direction recorded in the motor cortex by Georgopoulos and his colleagues. Later stages, such as the transformation from kinematic to dynamic parameters, or to levels of muscle activation, are beyond the scope of the present model. (shrink)

Contrary to Müller's claims, and in support of modular theories, genetic factors play a substantial and significant role in language. The finding that some children with specific language impairment (SLI) have nonlinguistic impairments may reflect improper diagnosis of SLI or impairments that are secondary to linguistic impairments. Thus, such findings do not argue against the modularity thesis. The lexical/functional distinction appears to be innate and specifically linguistic and could be instantiated in either symbolic or connectionist systems.

Müller argues that general cognitive skills and linguistic skills are not necessarily independent. However, cross-linguistic evidence from an inherited specific language disorder affecting productive rules suggests significant degrees of modularity, innateness, and universality of language. Confident claims about the overall nature of such a complex system still await more interdisciplinary research.

Müller attempts to downplay cases of dissociation between language and cognition as evidence against the modularity of language. We review cases of associations between verbal and nonverbal abilities as further evidence against the notion of language as an autonomous subsystem. We also point out a discrepancy between his proposal of homologies between nonhuman primates' communication and human language and recent proposals on the evolution of language.

Naturally occurring speech contains only a limited amount of complex recursive structure, and this is reflected in the empirically documented difficulties that people experience when processing such structures. We present a connectionist model of human performance in processing recursive language structures. The model is trained on simple artificial languages. We find that the qualitative performance profile of the model matches human behavior, both on the relative difficulty of center‐embedding and cross‐dependency, and between the processing of these complex recursive structures and (...) right‐branching recursive constructions. We analyze how these differences in performance are reflected in the internal representations of the model by performing discriminant analyses on these representations both before and after training. Furthermore, we show how a network trained to process recursive structures can also generate such structures in a probabilistic fashion. This work suggests a novel explanation of people's limited recursive performance, without assuming the existence of a mentally represented competence grammar allowing unbounded recursion. (shrink)

I argue that examining the explanatory power of the hypothesis of extended cognition (HEC) offers a fruitful approach to the problem of cognitive system demarcation. Although in the discussions on HEC it has become common to refer to considerations of explanatory power as a means for assessing the plausibility of the extended cognition approach, to date no satisfying account of explanatory power has been presented in the literature. I suggest that the currently most prominent theory of explanation in the special (...) sciences, James Woodward’s contrastive-counterfactual theory, and an account of explanatory virtues building on that theory can be used to develop a systematic picture of cognitive system demarcation in the psychological sciences. A major difference between my differential influence (DI) account and most other theories of cognitive extension is the cognitive systems pluralism implied by my approach. By examining the explanatory power of competing traditions in psychological memory research, I conclude that internalist and externalist classificatory strategies are characterized by different profiles of explanatory virtues and should often be considered as complementary rather than competing approaches. This suggests a deflationary interpretation of HEC. (shrink)

We focus on Karmiloff-Smith's Representational redescription model, arguing that it poses some problems concerning the architecture of a redescribing system. To discuss the topic, we consider the implicit/explicit dichotomy and the relations between natur al language and the language of thought. We argue that the model regards how knowledge is employed rather than how it is represented in the system.

Purely parallel neural networks can model object recognition in brief displays – the same conditions under which illusory conjunctions have been demonstrated empirically. Correcting errors of illusory conjunction is the “tag-assignment” problem for a purely parallel processor: the problem of assigning a spatial tag to nonspatial features, feature combinations, and objects. This problem must be solved to model human object recognition over a longer time scale. Our model simulates both the parallel processes that may underlie illusory conjunctions and the serial (...) processes that may solve the tag-assignment problem in normal perception. One component of the model extracts pooled features and another provides attentional tags that correct illusory conjunctions. Our approach addresses two questions: How can objects be identified from simultaneously attended features in a parallel, distributed representation? How can the spatial selectional requirements of such an attentional process be met by a separation of pathways for spatial and nonspatial processing? Our analysis of these questions yields a neurally plausible simulation of tag assignment based on synchronizing feature processing activity in a spatial focus of attention. (shrink)